Towards the use of Graphical Processor Units (GPUs) in QM-MM simulations.

GONZALEZ LEBRERO MC; NITSCHE, MATÍAS

Quilmes

Congreso; 1er Congreso Argentino de Bioinformática y Biología Computacional.; 2010

Asociación Argentina de Bioinformática y Biología Computacional (A2B2C)

The mayor cost in a QM-MM simulation resides in the calculation of the electronic structure of the quantumregion. For this, the optimization of this part for the calculation shifts the limits in the accuracy and size ofthe system who can be studied.This work presents an implementation of electronic structure calculations based on the Density FunctionalTheory (DFT), specically designed to be executed on a graphics processor (GPU). The software MOLECOLEwas used as starting point of this development. Following the literature on this subject, the known ecientalgorithms are implemented. However, novel adaptations that take into account both the characteristics ofthe used hardware, and the particular types of molecular systems in which this work focuses (QM-MM), areintroduced.The appeal of such hardware is its great computational power, which is a product of the highly parallelizedarchitecture implemented on it. In calculations where this parallelism can be exploited, it is possible to obtainsignicant improvements in execution times.With the software developed in this work an implementation which is at par with the known Gaussian '03in terms of numerical quality, but with a reduction in computing time to about ten times (10x) is achieved, bycomparing to a serial execution, and to four times (4x) by comparing to a parallelized execution over four CPUcores. Other comparisons are also made between the original software Molecole, SIESTA and an implementationthat targets conventional processors (CPU), but analogous to the one developed for GPU (see gure 1).Finally, it is to be expected that the experience gained through solving the various problems encounteredwhen working on this type of hardware could be generalized to future projects. Moreover, the results not onlymotivate to advance with the present work, but also demonstrate the usefulness of these processors to speed upthese type of calculations and possibly other similar.